Memory bias for threat in generalized anxiety disorder: the potential importance of stimulus relevance.

Information processing models propose that anxious individuals are characterized by memory biases for mood-congruent threat information. However, evidence for memory biases in generalized anxiety disorder (GAD) has been mixed at best. Given the heterogeneity of concerns in GAD, previous use of nomothetic stimulus sets may have precluded detection of memory biases. Therefore, in order to guarantee the relevance of the stimuli used, in the current study individuals with GAD each individually selected words that were of personal relevance to them. Using these idiographically selected words with 23 individuals with DSM-IV GAD and 23 non-anxious controls (NACs), results showed an implicit memory bias for threat words in individuals with GAD compared with NACs. Furthermore, there was additional evidence that individuals with GAD may also be characterized by explicit memory bias for threat words. The magnitude of group differences for explicit recall of threat words was similar to those previously observed in panic disorder. Limitations and future directions are discussed.